Mono and difunctionalisation of chiral ferrocenyl bis-acetals. X-Ray crystal structure of bis-1,1′-[(2S, 4S)-(hydroxymethyl)-2-dioxane1,3]-ferrocene

Article abstract of DOI:10.1016/S0022-328X(98)00449-5

The chiral 1,1′-bis-acetals, bis-1,1′-[(2S, 4S)-(hydroxymethyl)-2-dioxane1,3]-ferrocene (3) and 1,1′-bis-1,1′-[(2S, 4S)-(methoxymethyl)-2-dioxane1,3]-ferrocene (4) were synthesized. (3) was crystallographically characterised. The ortholithiation of (4) was studied in various conditions. Fair yields of monosubstituted compounds could be obtained with a complete regioselectivity in favor of the 2 position but the diastereoselectivities were moderate (up to 35%). Some disubstituted compounds can be isolated but in low yields (up to 8%). The regioselectivity is complete in favor of the 2,2′-disubstituted isomer. Only the diastereoisomer with two opposite planar chiralities is observed.

Full text of DOI:10.1016/S0022-328X(98)00449-5

Journal of Organometallic Chemistry 565 (1998) 115–124
Mono and difunctionalisation of chiral ferrocenyl bis-acetals. X-Ray
crystal structure of bis-1,1%-[(2S,
4S)-(hydroxymethyl)-2-dioxane1,3]-ferrocene1
Gabriel Iftime, Jean-Claude Daran, Eric Manoury *, Gilbert G.A. Balavoine 2
Laboratoire de Chimie de Coordination, 205 route de Narbonne, F-31077 Toulouse Cedex, France
Received 27 November 1997
Abstract
The chiral 1,1%-bis-acetals, bis-1,1%-[(2S, 4S)-(hydroxymethyl)-2-dioxane1,3]-ferrocene (3) and 1,1%-bis-1,1%-[(2S, 4S)-(methoxy-
methyl)-2-dioxane1,3]-ferrocene (4) were synthesized. (3) was crystallographically characterised. The ortholithiation of (4) was
studied in various conditions. Fair yields of monosubstituted compounds could be obtained with a complete regioselectivity in
favor of the 2 position but the diastereoselectivities were moderate (up to 35%). Some disubstituted compounds can be isolated
but in low yields (up to 8%). The regioselectivity is complete in favor of the 2,2%-disubstituted isomer. Only the diastereoisomer
with two opposite planar chiralities is observed. © 1998 Elsevier Science S.A. All rights reserved.
Keywords: Iron; Ferrocene; Regioselectivity; Diastereoselectivity; Chirality
1. Introduction
chiral directing groups were described such as an amine
([2]c), an acetal [3], a sulfoxide [4] or an oxazoline [5].
Only very recently were ‘direct methods’ described
which use an achiral ferrocene derivative but a chiral
base [6] for the ortholithiation, sometimes with very
high enantioselectivities (up to 99% ee) ([6]b).
The enantioselective synthesis of C2-symmetric-tetra-
substituted ferrocene derivatives with planar chirality
only recently started to attract more attention, despite
the attractive structure of this type of compound. Most
of the reported methods for their synthesis followed the
diastereoselective ortholithiation strategy applied to fer-
rocenes bearing one chiral directing group on each
cyclopentadienyl ring. This chiral directing group may
be an amino group [7], an oxazoline [8]. Only very
recently was reported enantioselective syntheses of C2-
symmetric-tetrasubstituted ferrocenes from 1,1%-fer-
rocenecarboxaldehyde with a bis-aminoalkoxide as an
intermediate [9] or from an 1,1%-ferrocenyl diamides
with nBuLi/sparteine [10]. Nevertheless, C2-symmetric
Owing to their use in many fields for example organic
synthesis, supramolecular chemistry, materials chem-
istry or catalysis, ferrocenes with planar chirality are of
high interest [1]. A number of years ago, the procedure
reported by Ugi and coworkers concerning the
diastereoselective ortholithiation of (R)-N,N-dimethyl-
1-ferrocenyl ethyl amine (92% of diastereoisomeric ex-
cess) became the method of choice for introducing
planar chirality on a ferrocene ([2]a,b). More recently,
many new efficient methods have been reported. Most
of them used ferrocene derivatives which possess a
chiral group bearing heteroatom acting as a directing
group for diastereoselective ortholithiation. Various
* Corresponding author. Fax: +33 56 1333131; e-mail:
manoury@lcc-toulouse.fr or balavoin@lcc-toulouse.fr
1 Dedicated to Professor Michael Bruce on the occasion of his 60th
birthday in recoginition of his important contribution to
organometallic chemistry.
2 Also corresponding author.
0022-328X/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved.
PII S0022-328X(98)00449-5

116
G. Iftime et al. / Journal of Organometallic Chemistry 565 (1998) 115–124
2. Results and discussion
2.1. Synthesis of the acetals (3) and (4)
The synthesis of (3) was similar to that used by
Kagan et al. to obtain (1) (see Scheme 2). 1,1%-fer-
rocenecarboxaldehyde is heated in neat trimethylortho-
formate in the presence of a catalytic amount of
paratoluenesulfonic acid to obtain the bis dimethylac-
etal. Transacetalization of the crude bis-acetal is carried
out in chloroform in the presence of commercially
Scheme 1.
available (S)-(–)-1,2,4-butanetriol and
a
catalytic
amount of paratoluenesulfonic acid. The desired bis-ac-
etal is obtained as a mixture of diastereoisomers. After
recrystallisation in toluene, pure (3) is obtained as a
single diastereoisomer. (4) is synthesized from (3) by
deprotonation of the hydroxy group using NaH fol-
lowed by alkylation with iodomethane.
tetrasubstituted ferrocene have already found successful
application to asymmetric catalysis, namely to palla-
dium-catalyzed asymmetric cross coupling (ee up to
93%) ([7]a) and to asymmetric allylic substitution (ee up
to 99%) [11].
2.2. X-ray crystal structure of (3)
In 1993, Kagan et al. showed that the monoacetal (1)
(see Scheme 1) ([3]a) allows the synthesis of various
2-substituted ferrocenecarboxaldehydes with good
yields and in an essentially enantiomerically pure form.
Our idea was to consider the corresponding bis-acetal
(4) with one acetal group on each cyclopentadienyl ring
as starting materials to obtain C2-symmetric 2,2%-disub-
stituted 1,1%-ferrocenedicarboxaldehyde. This type of
molecule, owing to the large reactivity of the aldehyde
function, can be an intermediate in the synthesis of
numerous compounds of great interest in various fields
of chemistry. Here we report the synthesis of chiral bis
acetal bis-1,1%-[(2S, 4S)-(hydroxymethyl)-2-dioxane1,3]-
ferrocene (3) and bis-1,1%-[(2S, 4S)-(methoxymethyl)-2-
dioxane1,3]-ferrocene (4) and our results with regard to
the reactivity and stereochemistry of the mono- and
dilithiation of (4) under various conditions (see Scheme
3). (3) was identified by X-ray analysis.
The molecular structure of (3) shown in Fig. 1, is
essentially as expected and confirms the formulation of
the complex. Bond lengths and angles are collected in
Table 1. As this complex is enantiomerically pure and
crystallizes in the noncentrosymmetric P212121 space
group, its absolute configuration was determined by
refining the Flack enantiopole parameter x [12], which
is defined as
F20=(1–x)F(h)2+xF(−h)2
x is then the fractional contribution of F(−h) to the
observed structure amplitude, and it is sensitive to the
polarity of the structure. The x value, 0.01(3), agrees
with the absolute configuration expected from the syn-
thetic route for the C(14) and C(64) atoms. The two
acetal carbon C(11) and C(61) are asymmetric and they
have the same absolute configuration S. It is worth
Scheme 2.

G. Iftime et al. / Journal of Organometallic Chemistry 565 (1998) 115–124
117
Fig. 1. Cameron view of the molecule with ellipsoids plotted at the 50% probability level. Hydrogen atom but those attached to the hydroxy group
have been omitted for clarity.
pointing out that this S configuration is identical to the
one we recently reported for a ferrocenophane acetal
complex [13] and confirms the configuration of the
acetal carbon and the planar chirality which was estab-
lished by Kagan et al. ([3]a).
The ferrocene moiety reveals the expected geometry
with no unusual bond lengths or angles. The two Cp
rings which are eclipsed, are virtually parallel making a
dihedral angle of 0.6°.
Table 1
˚
Bond lengths (A) and bond angles (°) with e.s.d.s in parentheses.
The two dioxane rings have a slightly distorted chair
conformation. The four atoms, O(1), O(2), C(12), C(14)
or O(3), O(4), C(62), C(64) are coplanar while atoms
O(1)ꢀC(11)
O(1)ꢀC(14)
O(2)ꢀC(11)
O(2)ꢀC(12)
O(3)ꢀC(61)
O(3)ꢀC(64)
O(4)ꢀC(61)
O(4)ꢀC(62)
O(15)ꢀC(15)
O(65)ꢀC(65)
1.426(6)
1.440(6)
1.390(7)
1.429(7)
1.409(6)
1.432(6)
1.410(6)
1.437(8)
1.417(8)
1.409(7)
C(1)ꢀC(11)
C(6)ꢀC(61)
C(7)ꢀC(8)
C(8)ꢀC(9)
C(9)ꢀC(10)
C(12)ꢀC(13)
C(13)ꢀC(14)
C(14)ꢀC(15)
C(62)ꢀC(63)
C(63)ꢀC(64)
C(64)ꢀC(65)
O(65)ꢀH(65)
O(65)· · ·O(1)
O(1)· · ·H(65)
O(1)ꢀC(14)ꢀC(13)
O(1)ꢀC(14)ꢀC(15)
C(13)ꢀC(14)ꢀC(15)
O(15)ꢀC(15)ꢀC(14)
O(3)ꢀC(61)ꢀO(4)
O(3)ꢀC(61)ꢀC(6)
O(4)ꢀC(61)ꢀC(6)
O(4)ꢀC(62)ꢀC(63)
C(62)ꢀC(63)ꢀC(64)
O(3)ꢀC(64)ꢀC(63)
O(3)ꢀC(64)ꢀC(65)
C(63)ꢀC(64)ꢀC(65)
O(65)ꢀC(65)ꢀC(64)
1.485(7)
1.461(7)
1.405(7)
1.432(8)
1.398(8)
1.507(9)
1.511(9)
1.491(8)
1.504(9)
1.508(8)
1.504(7)
0.94(4)
˚
C(11) or C(61) lie 0.648 or 0.625 A below and C(61) or
˚
C(62) 0.639 or 0.664 A above these planes, respectively.
The arrangement of these two dioxane rings with re-
spect to the Cp rings are different. The dioxane at-
tached to C(1) makes a dihedral angle of 48.1° with the
corresponding Cp ring whereas the other dioxane is
roughly coplanar with the Cp ring making a dihedral
angle of only 10.2°. This difference may be related to
the occurrence of intermolecular hydrogen bonds which
engage the O(1) atom of the first dioxane ring.
O(15)ꢀH(15)
O(15)· · ·O(65)
O(65)· · ·H(15)
0.99(4)
2.753(6)
1.80(4)
2.771(5)
1.85(4)
One of the interesting feature of the structure is
indeed the occurrence of a complex system of inter-
C(11)ꢀO(1)ꢀC(14)
C(11)ꢀO(2)ꢀC(12)
C(61)ꢀO(3)ꢀC(64)
C(61)ꢀO(4)ꢀC(62)
C(2)ꢀC(1)ꢀC(11)
C(5)ꢀC(1)ꢀC(11)
C(7)ꢀC(6)ꢀC(61)
C(10)ꢀC(6)ꢀC(61)
C(6)ꢀC(10)ꢀC(9)
O(1)ꢀC(11)ꢀO(2)
O(1)ꢀC(11)ꢀC(1)
O(2)ꢀC(11)ꢀC(1)
O(2)ꢀC(12)ꢀC(13)
C(12)ꢀC(13)ꢀC(14)
111.8(4)
111.2(4)
113.0(4)
110.9(4)
126.7(5)
125.1(5)
127.3(4)
125.9(5)
108.8(5)
111.6(4)
108.1(4)
109.4(4)
110.6(4)
110.6(5)
108.7(4)
107.3(4)
114.5(5)
106.2(5)
112.1(4)
109.6(4)
110.0(4)
109.8(5)
109.5(5)
109.7(4)
104.9(4)
113.6(5)
108.1(4)
Fig. 2. Stereoview of the packing showing intermolecular hydrogen
bonding. Dotted lines indicate O–H· · ·OH and O–H· · ·O hydrogen
bonds. For clarity, only half of the packing is represented.
O(15)ꢀH(15)· · ·O(65) 161(6)
O(65)ꢀH(65)· · ·O(1) 145(2)

118
G. Iftime et al. / Journal of Organometallic Chemistry 565 (1998) 115–124
Fig. 3. Packing diagram viewed down the b axis and showing the wave-like layer built by the intermolecular hydrogen bonds.
molecular hydrogen bonds between the molecules to
form a two-dimensional network (Fig. 2). One of the
hydroxy group O(65) acts as both a donor and an
acceptor for intermolecular hydrogen bonds whereas
the other hydroxy O(15) acts as only a donor. More-
over, the oxygen atom O(1) of one of the 1,3-dioxane
ring is also an intermolecular hydrogen-bond acceptor.
Firstly, molecules are linked by O–H· · ·OH hydrogen
bonds via unit cell translations to form infinite chain in
the diagonal xy direction [O(15)· · ·O(65){x–1,1+y,z}
after electrophilic quenching of the lithiated intermedi-
ates by Bu3SnCl to give (5a) and (6a) or Me3SiCl to
yield (5b) and (6b) (see Table 2 and Scheme 3).
The yields of the monosubstituted bisacetals (5)
formed by deprotonation with tBuLi was not higher
than 22% (entry 1, Table 2) despite many attempts in
various conditions. The yields of disubstituted bisac-
etals (6) were even lower. We supposed that the ap-
proach to the ferrocenyl protons was very hindered by
the two large acetal groups. Using a less bulky base
might be a way to deprotonate (4) more efficiently. The
yields of (5) when MeLi was employed as base were
increased (up to 69%, isolated yield) but not the yields
of (6) (entries 6–9, Table 2). With 1 equiv. of MeLi, no
deprotonation occurred. The deprotonation started
only when more than 1 equiv. of MeLi was used
(entries 6, 7, 8, Table 2).
In every case, (5) ((5a) or (5b)) is a mixture, in
various proportions, of only two diasteroisomers,
amongst the four possible ones which can be synthe-
sized by monofunctionalisation of (4) in a 2 or 3
position. The hydrolysis of each mixture of
diastereoisomers ((5a) or (5b)) provides only one alde-
˚
2.753(6) A]. Secondly, two chains related by the sym-
metry operator 1/2+x,1/2−y,−z are also linked by
O–H· · ·O hydrogen bonds which develop between the
hydroxy group O(65) and one of the 1,3-dioxane ring
˚
[O(65)· · ·O(1) {1/2+x,1/2−y,−z}, 2.771(5) A]. This
second link between chains results in the build up of
wave-like layers which develop perpendicular to the c
direction (Fig. 3).
2.3. Deprotonation of (4)
The deprotonation of (4) was studied in different
conditions by characterization of the reaction mixture

G. Iftime et al. / Journal of Organometallic Chemistry 565 (1998) 115–124
119
Table 2
Functionalisation of the bis-acetal (4)
Entry
Conditions for the deprotonation step
Electrophile
Yield in (5)a (de, major isomerb)
Yield in (6)a
1
2
3
4
5
6
7
8
9
2.1 eq. tBuLi in Et2O; −78°C then RT, 2 h
2.1 eq. TBuLi in THF; −78°C then RT, 2 h
4.2 eq. TBuLi in Et2O; −78°C then RT, 2h
2.1 eq. TBuLi in Et2O; reflux, 15 h
1.05 eq. tBuLi in Et2O; −78°C then RT, 2 h
2.1 eq. MeLi in Et2O; −78°C then RT, 2 h
1.6 eq. MeLi in Et2O; −78°C then RT, 2 h
1.05 eq. MeLi in Et2O; −78°C then RT, 2 h
2.1 eq. MeLi in Et2O; −78°C then RT, 2 h
Bu3SnCl
Bu3SnCl
Bu3SnCl
Bu3SnCl
Bu3SnCl
Bu3SnCl
Bu3SnCl
Bu3SnCl
Me3SiCl
(5a) 22% (19%, S)
(5a) 0%c
(6a) 8%
(6a) 0%c
(6a) 0%d
(6a) 0%e
(6a) 0%c
(6a) 5%
(6a) 2%
(6a) 0%c
(6b) 5%f
(5a) 0%d
(5a) 0%e
(5a) 0%c
(5a) 69% (34%, R)
(5a) 44% (33%, R)
(5a) 0%c
(5b) 64% (36%, R)
a Isolated yields.
b The given configuration is related to the planar chirality.
c Starting materials recovered.
d Not isolated from a complex mixture.
e Starting materials completely decomposed.
f Isolated only after hydrolysis to the dialdehyde (6b).
hyde ((7a) or (7b)). If the functionalisation occurred in
a 2 position, the H-NMR spectrum must show for the
of ca. 1.3 Hz, and two similar doublets of doublets. The
first case is observed, for (7a) and (7b), so they are the
2-substituted 1,1%-ferrocenecarboxaldehydes and there-
fore (5) is a mixture of (5)A and (5)B (see Scheme 4).
1
H on the disubstituted cyclopentadienyl ring, a triplet
with a coupling constant of ca. 2.5 Hz, and two dou-
blets of doublets with one coupling constant of ca. 2.5
Hz and the other one of ca. 1.3 Hz [14]. If the function-
Furthermore,
(S)-2-(trimethylsilyl)-1,1%-ferrocenecar-
boxaldehyde has already been synthesized by another
pathway [13] and all the physical data (except rotarory
power) are the same for (S)-2-(trimethylsilyl)-1,1%-fer-
rocenecarboxaldehyde and (7b), confirming the 2-sub-
stitution in (7b). The integration of the different acetal
signals in the 1H-NMR spectrum of (5a) or (5b) gave us
the diastereoisomeric excess of the different mixtures of
diastereoisomers. In the case of (7b), the value and the
sign of the rotarory power gave the enantiomeric excess
of (7b) and the absolute configuration of the major
enantiomer. So, we can identify (5b)A and (5b)B. Mak-
ing the very logical hypothesis that the stereoselectivity
of the functionalisation is made at the lithiation step
and therefore does not depend of the electrophile, (5a)A
and (5a)B were identified. The measurement of the ee of
(7b) was not precise because it is difficult to obtain it
1
alisation occurred in a 3 position, the H-NMR spec-
trum must show for the H on the disubstituted
cyclopentadienyl ring, a triplet with a coupling constant
Scheme 3.
Scheme 4.

120
G. Iftime et al. / Journal of Organometallic Chemistry 565 (1998) 115–124
Table 3
de of (5b) by integration of the acetal
signals (%)
ee of (7b)by measure- ee of (9b) by measurement ee of (9b) using
ee of (10b) by measure-
ment of [h]D
ment of [h]D (%)
of [h]D (%)
Eu(hfc)3 (%)
36
ca. 25
ca. 25
35
31
very pure. Therefore, we decided to reduce (7b) into
(9b) (see Scheme 3) and afterwards to transform (9b)
into (10b). The corresponding diol and ferrocenophane
are known in an enantiomerically pure form [13], so the
measurement of the rotarory power of (9b) and (10b)
can give values of the ee of (7b). Furthermore, the diol
have been resolved in 1H-NMR using Eu(hfc)3 as a
chiral chemical shift reagent. The measurement of the
rotarory power of (10b) is not very precise because of
the low value of [h]D (36.2) and it is even worse for (9b)
([h]D= +8.17 (CHCl3, c=0.93)). The data concerning
the diastereoisomeric excess of (5b) and the enan-
tiomeric excess of (7b), (9b) and (10b) are collected in
Table 3. All this data are well related and support our
hypothesis. With tBuLi (Table 2, entry 1), the major
diastereoisomer have a S configuration like the alde-
hydes (2) obtained by deprotonation of (1) (see Scheme
1) but the selectivity is much lower (19% instead of
more than 98%). Moreover, using MeLi, the
diastereoselectivity is also low (de=35%) but the major
diastereoisomer have now a R configuration, although
tion and the other one in a 3 position. Furthermore, the
signal for one H on the cyclopentadienyl ring is a triplet
with a coupling constant of 2.5 Hz, indicating that this
hydrogen atom have two neighboring hydrogen atoms
in a 2 position related to it [14]. So each Cp ring of (8a)
is 1,2 disubstituted, therefore each Cp ring of (6a) is
also 1,2 disubstitued. (6a) is (6a)A (see Scheme 4) and
(8a) is the meso 2,2%-disubstituted 1,1%-ferrocenedicar-
boxaldehyde. When Me3SiCl was used as an elec-
trophile, the dialdehyde (8b) was isolated. According to
1
its H-NMR spectrum, (8b) is a 2,2%-disubstituted 1,1%-
ferrocenedicarboxaldehyde which is not the dl
diastereoisomer (the (R,R)-2,2%-bis (trimethylsilyl) 1,1%-
ferrocenecarboxaldehyde was synthesized by an other
route (see Scheme 5); the structure and stereochemistry
of this compound was confirmed by X-ray crystal anal-
ysis [9]). Once again, (8b) is the meso diastereoisomer.
Finally, the meso structure of (8a) and (8b) was confi-
rmed by their rotatory power which was found to be 0.
2.4. Conclusion
the deprotonation of (1) yields mainly the
S
diastereoisomer (de=50%, see Table 4). (3) and (1) do
not react in a similar way.
The reactivity of (3) is not directly related to that of
(1). The two acetal groups do not act independently.
They are close enough to interact together with the
tBuLi or MeLi delivering a completely new system.
But, unfortunately, although the regioselectivities are
very high in all cases, the diastereoselectivity of the
synthesis of (5) is quite low (19% with tBuLi; 35% with
MeLi). Furthermore, the yields of the disubstituted (6)
are very low and only the achiral meso stereoisomer
which is not useful for the synthesis of chiral tetrasub-
stituted ferrocenes can be observed. But it is worth to
point out that the difunctionalisation of (4) occurs with
a perfect regio and diastereoselectivity.
In any of the cases examined, the yields of tetrasub-
stituted (6) (see Scheme 3) are low (B8%), even with
an excess of MeLi, but only one diastereoisomer was
observed amongst the ten possible compounds which
can be obtained by a double deprotonation in a 2 or 3
1
position. The H-NMR spectrum of (6a) shows that the
two Cp rings and the two acetal functions are not
equivalent, so there is no C2 axis in this molecule. After
hydrolysis, (6a) yields only one disubstituted 1,1%-fer-
1
rocenedicarboxaldehyde (8a). The H-NMR spectrum
of this compound shows the equivalence of the Cp rings
and of the two aldehyde functions, therefore (8a) must
be the 2,2% or 3,3% functionalised compound, ruling out
all stereoisomers with one functionalisation in a 2 posi-
Table 4
Trimethylsilylation of the monoacetal (1)
Entry
Base
Yield in (10) (%)
de of (10), major isomera
1
2
tBuLi
MeLi
93
12
\98%, Sb
50%, S
a In the experimental conditions described in ref. ([3]a) the given
configuration is related to the planar chirality.
b From reference ([3]a).
Scheme 5.

G. Iftime et al. / Journal of Organometallic Chemistry 565 (1998) 115–124
121
3. Experimental section
CH).[h]D= −18.7(CHCl3, c=0.60) MS (IE, 70 eV)
m/e: 418 (M, 33%), 75 (73%), 57 (64%), 45 (77%), 43
(50%), 31 (100%). Anal. Found: C, 57.51; H, 6.44.
C20H26FeO6 Calc.: C, 57.44; H, 6.27.
3.1. General procedure and reagent syntheses
All of the reactions were carried out in the absence of
air using standard Schlenk techniques and vacuum-line
manipulations. Chlorotrimethylsilane and iodomethane
3.3. X-ray crystal determination of (3)
˚
was distilled on calcium hydride and stored on 3–4 A
The data were collected at r.t. (293 K) on a Stoe
Imaging Plate Diffraction System (IPDS). The crystal-
to-detector distance was 60 mm. 143 exposures (4 min.
per exposure) were obtained with 0B€B200° and
with the crystals rotated through 1.4° in €. Coverage of
the unique set was over 99% complete to at least 28° in
q. Crystal decay was monitored by measuring 200
reflections per image. The final unit cell parameters
were obtained by the least-squares refinement of 2000
reflections. Only statistical fluctuations were observed
in the intensity monitors over the course of the data
collection. Owing to the rather low mx value, 0.16, no
absorption correction was considered.
molecular sieves. Bu3SnCl was distilled under reduced
pressure on carborandum and kept on 3–4 A molecular
˚
sieves. 1,1%-ferrocenedicarbaldehyde was synthesized ac-
cording to ref. [15]. Other compounds were used as
commercial samples. All solvents were dried before use.
Thin layer chromatography was carried out on Merck
Kieselgel 60F254 precoated silicagel plates. Preparative
flash chromatography was performed on Merck
Kieselgel. Instrumentation: Bruker AM250 (1H, 13C),
Hewlett-Packard HP MSD 7590 (GC/MS), Enraf-Non-
ius CAD4 (X-ray). Elemental analyses were performed
by the Service d’Analyse du Laboratoire de Chimie de
Coordination, Toulouse (France).
The structure was solved by direct methods (SIR92)
[16] and refined by least-squares procedures on Fo. H
atoms were located on difference Fouriers maps, but
those attached to C atoms were introduced in calcula-
3.2. Synthesis of bis-1,1%-[(2S, 4S)-
(hydroxymethyl)-2-dioxane-1,3]-ferrocene (3)
˚
tion in idealized positions(d(CH)=0.96 A) and their
A mass of 6.0 g (24.6 mmol) of 1,1%-ferrocenedicar-
boxaldehyde was heated overnight at 80°C in 100 ml of
trimethyl orthoformate together with 0.5 g (2.63 mmol)
of paratoluenesulfonic acid monohydrate. Anhydrous
potassium carbonate was then added and the mixture
was stirred one more hour at 80°C. After cooling, the
solution was diluted with ether, then filtered on celite.
After evaporation of the solvents, a brown oil was
obtained.
atomic coordinates were recalculated after each cycle.
They were given isotropic thermal parameters 20%
higher than those of the carbon to which they are
attached. The coordinates of the H atoms attached to
the O atoms were refined with a fixed isotropic thermal
parameter of 0.07. Least-squares refinements were car-
ried out by minimizing the function Sw(ꢀFoꢀ−ꢀFcꢀ)2,
where Fo and Fc are the observed and calculated struc-
ture factors. The weighting scheme used in the last
refinement cycles was w=w%[1−{(DF/6|(Fo)}2]2 where
w%=1/Sn1ArTr(x) with three coefficients Ar for the
Chebyshev polynomial ArTr(x) where x was Fc/Fc(max)
[17]. Models reached convergence with R=S(ꢀꢀFoꢀ−
ꢀFcꢀꢀ)/S(ꢀFoꢀ) and Rw=[Sw(ꢀFoꢀ−ꢀFcꢀ)2/Sw(Fo)2]1/2, hav-
ing values listed in Table 4. Criteria for a satisfactory
complete analysis were the ratios of rms shift to stan-
dard deviation less than 0.1 and no significant features
in final difference maps. Details of data collection and
refinement are given in Table 5.
To a solution of 6.0 g (56 mmol) of (S)-(–)-1,2,4-bu-
˚
tanetriol in 30 ml of chloroform, was added 4 A
molecular sieves and 0.5 g (2.63 mmol) of paratoluene-
sulfonic acid monohydrate. After 5 min stirring at r.t.,
was added a solution of the brown oil in 30 ml of
chloroform and the mixture was heated at 60°C
overnight. Anhydrous potassium carbonate was then
added and the mixture was cooled down back to r.t.
The mixture was then filtered on celite, which was
washed by several fractions of dichloromethane. The
organic solution was evaporated with a rotavapor to
yield a brown oil. The crude materials was chro-
matographed on silicagel with ethyl acetate to yield
The calculations were carried out with the CRYS-
TALS package programs [18] running on a PC. The
drawing of the molecule was realized with the help of
CAMERON [19].
1
6.27 g of a yellow solid (61%). H–NMR (l (ppm),
CDCl3): 5.40 (s, 2H, O–CH
4H, Cp), 4.23–4.08 (m, 6H, 4H on Cp and 2H –CH
O), 4.0–3.85 (m, 4H, CH–O), 3.63 (m, 4H, –CH2–O),
2.72 (s, 2H, –OH), 1.78 (br. qd., J(CH2gem)=13 Hz
and 5.1 Hz, 2H, –CH2–), 1.41 (dd, J(CH2gem)=13 Hz,
and J=1.2 Hz, 2H, –CH2
–).13C–NMR (l (ppm),
CDCl3): 99.97 (O–CH–O), 86.31, 77.37, 68.67, 68.64,
67.77, 67.60, 66.46, 65.58, 26.78(–CH2–CH2–
6 –O), 4.30 (t, J=1.8 Hz,
6
–
6
6
3.3.1. Supplementary materials a6ailable
The fractional atomic coordinates and the an-
isotropic thermal parameters for non hydrogen atoms
have been deposited at the Cambridge Crystallographic
Data Center.
6
6
6

122
G. Iftime et al. / Journal of Organometallic Chemistry 565 (1998) 115–124
Table 5
Crystal data
addition of pentane yield 1.02 g of an orange oil(2.28
mmol, 95%).
1H-NMR (l (ppm), CDCl3): 5.38 (s, 2H, O–CH
4.29 (m, 4H, Cp), 4.19 (ddd, J=10.2 Hz, 5.0 and 1.0
Hz, 2H, CH–O), 4.10 (t, J=1.8 Hz, 4H, Cp), 3.98 (m,
2H, CH–O), 3.88 (ddd, J=11.3, 10.2 and 2.6 Hz, 2H,
CH–O), 3.51 (dd, J=11 and 6.2 Hz, 2H, CH2–OMe),
3.38 (dd, J=10.2 and 4.8 Hz, 2H, CH2–OMe, ABX
with H with l at 3.51 and 3.98 ppm), 3.39 (s, 6H, CH3),
1.74 (br dq, J=ca. 12 Hz (with H at 3.98, 3.88 and
1.45 ppm) and 5.1 Hz (with H at 4.19 ppm), 2H, CH2),
1.45 (br. d, J(CH2gem)=11.3 Hz, 2H, CH2)
.13C-NMR
(l (ppm), CDCl3): 99.87(O–CH–O), 86.22, 75.86,
75.54, 68.76, 67.46, 66.59, 59.28(O–CH3), 27.92(–
CH2–CH2–CH). [h]D= −49.0 (CHCl3, c=0.45) GC-
6
–O),
Crystal parameters
Compound
Formula weight (g)
Shape (color)
Size (mm)
Crystal system
Space group
C20H26O6Fe
418.27
Box (yellow)
0.23, 0.20, 0.19
Orthorhombic
P212121
6
6
6
6
6
6
Unit cell dimensions
˚
a (A)
5.8730(4)
14.363(2)
22.592(1)
93.66(1)
1905.7(4)
4
881
1.458
8.212
˚
b (A)
6
˚
c (A)
6
i (°)
6
3
˚
V (A )
6
Z
F(000)
6
zcalcd (g cm−3
)
MS (IE, 70 eV) m/e: 448 (10%), 447 (36%), 446 (M,
100%), 149 (12%), 135 (10%), 121 (14%), 185 (12%).
Anal. Found: C, 59.22; H, 7.29. C22H30FeO6. Calc.: C,
59.20; H, 6.77.
v(Mo–Kh) (cm−1
)
Data collection
Diffractometer
Monochromator
Radiation
Scan mode
Detector distance (mm)
€ Range (°)
IPDS Stoe
Graphite
Mo–Kh (u=0.71073)
€
3.5. Functionalisation of (4) (general procedure):
synthesis of (5a) and (6a)
60
0B€ B200.2
€ Incr. (°)
Exposure time (min)
2q range (°)
No. of reflections collected
No. of unique reflections (Rint
Reflections used (I\2|(I))
1.4
4
A mass of 0.15 g (0.33 mmol) of bisacetal (4) was
dissolved in 10 ml of dry ether in a schlenk tube, then
the solution was cooled down to −78°C. 470 ml of a
1.5 M solution of t-BuLi in pentane (7.1 mmol, 2.1 eq.)
was added and the mixture was stirred 15 min. at
−78°C, then 2 h at r.t. A yellow precipitate appeared.
The mixture was cooled down again to −78°C and
tributyltin chloride (4 eq.) was added. After 15 min.
stirring at −78°C, the mixture was allowed to come
back at r.t. After 1 h stirring at r.t., the excess of
electrophile was destroyed by addition of 0.5 ml of
isopropylamine first, afterwards by water. The solution
was extracted with ether, washed with water, dried on
magnesium sulfate, evaporated to yield an orange oil
which was chromatographed on silicagel with a cyclo-
hexane/ ether mixture, to yield 31 mg of (6a) (8%), 167
mg of (5a) (69%) as a mixture of diastereoisomer (ratio
(5a)A/(5a)B=1.5; de=19%) and 14 mg of 1,1%-fer-
rocenedicarboxaldehyde (12%).
10.0B2qB56.1
7190
4423(0.063)
2706
)
Refinement
R
Rw
Weighting scheme
Coefficient Ar
GOF
0.0512
0.0575
Chebyshev
3.60, −2.11, 2.56
1.090
0.0262
−0.563/0.466
0.01(3)
(D/|)max
Dzmin/Dzmax
Flack’s parameter
Ls parameters
252
3.4. Synthesis of bis-1,1%-[(2S,
4S)-(methoxymethyl)-2-dioxane-1,3]-ferrocene (4)
In a schlenk tube, 330 mg of a 60% dispersion of
NaH in mineral oil (8.33 mmol) was washed with two
portions of dry cyclohexane. Then 5 ml of dry THF
was added. To the hydride suspension, cooled down to
0°C, was added a solution of 1 g of bisacetal (3) in 30
ml of dry THF. After 30 min stirring at 0°C, was added
0.54 ml (8.82 mmol) of iodomethane. After one night
stirring at r.t., the excess of hydride was destroyed by
addition of ethanol followed by water. The solution
was extracted with ether, washed with water and dried
on magnesium sulfate. After evaporation of the solvent
with a rotavapor, the crude materials was chro-
matographed on silicagel with ether and finally precipi-
tated by dissolution in ether followed by a slow
(5a):1H-NMR (l (ppm), CDCl3): 5.40 (s, 1H (5a)A,
O–CH–O), 5.39 (s, 1H (5a)B, O–CH–O), 5.37 (s, 1H
(5a)B, O–CH–O), 5.28 (s, 1H (5a)A, O–CH–O), 4.48
(m, 1H (5a)A+(5a)B, Cp), 4.32–4.07 (m., 5H (5a)A+
(5a)B), 4.03–3.78 (m 7H (5a)A+(5a)B), 3.67–3.29 (m,
10H (5a)A+(5a)B), 1.75 (m, 2H (5a)A+(5a)B, –CH–),
1.52 (m, 6H (5a)A+(5a)B, SnBu3), 1.34 (m, 8H (5a)A+
(5a)B, SnBu3+ –CH–), 1.00 (m, 6H (5a)A+(5a)B,
SnBu3), 0.89 (m, 9H (5a)A+(5a)B, Sn(CH2)3CH6 3).
Anal. Found: C, 55.67; H, 7.96. C34H56FeO6Sn Calc.:
C, 55.58; H, 7.62.
1
(6a): H-NMR (l (ppm), CDCl3): 5.33 (s, 1H, O–
CH–O), 5.26 (s, 1H, O–CH–O), 4.47 (dd, J=2.2 and
1.1 Hz, 1H, C5H3), 4.45 (dd, J=2.2 and 1.1 Hz, 1H,

G. Iftime et al. / Journal of Organometallic Chemistry 565 (1998) 115–124
123
C5H3), 4.19 (m, 2H, 1H C5H3 and 1H CH–O), 4.12 (m,
2H, 1H C5H3 and 1H, CH–O), 3.89 (m, 6H, 2H C5H3
and 4H CH–O), 3.50 (dd, J=9.9 and 5.5 Hz, 2H,
–CH–O), 3.36 (m, 8H, 2×3H OCH3 (s, 3H each at
3.37 and 3.35 ppm) and 2H –CH2–O), 1.75–0.8 (m,
58H, 4H –CH– and 54H –SnBu3). 13C-NMR (l
(ppm), CDCl3): 100.98 (O–CH–O), 100.90 (O–CH–
O), 91.02, 90.85, 76.28, 76.12, 75.82, 75.75, 75.39, 72.00,
71.79, 69.75, 69.64, 68.71, 68.34, 66.57, 66.39, 59.43
(OCH3), 59.29 (OCH3), 29.31 (SnBu3), 28.57 (O–
84.36, 80.57, 76.67, 74.65, 73.81, 29.22 (SnBu3), 27.38
(SnBu3), 13.69 (SnBu3), 10.94 (SnBu3). [h]D=0.0
(CHCl3, c=0.15)
(8b): Obtained from an impure (4b). (6b) was purified
by preparative layer chromatography on silicagel with a
cyclohexane/ether mixture. 1H-NMR (l (ppm), CDCl3):
9.96 (s, 2H, CHO), 5.00 (m, 2H, Cp), 4.76 (t, J=2.2
Hz, 2H, Cp), 4.59 (m, 2H, Cp), 0.32 (s, 18H, SiMe3).
GC-MS: m/e: 388 (M+2, 12%), 387 (M+1, 32%), 386
(M, 100%), 371 (46%), 32 (26%). [h]D=0.0 (CHCl3,
c=0.12)
CH2C
6
H2–CH–O), 28.30 (O–CH2C
6
H2–CH–O), 27.57
H3).
(SnBu3), 13.72, (SnBu3), 10.68 (Sn(CH2)3C
6
3.8. Reduction of the bisaldehyde (7b): synthesis of (9b)
3.6. Synthesis of (5b) and (6b)
In a Schlenk tube under argon, to a solution of
dialdehyde (7b) in methanol (ca. 0.05 mol l−1), cooled
at 0°C, was added a solution of NaBH4 (15 eq.) in
aqueous 2N sodium hydroxide solution. The mixture
was stirred overnight at r.t., then the methanol was
evaporated with a high vacuum pump. The organic
phase was extracted with ether, washed with water,
dried on magnesium sulfate and evaporated. The crude
product was purified by flash chromatography on sil-
icagel to yield pure (9b) (79% yield).
Experimental conditions similar to Section 3.3 with
MeLi as a base and Me3SiCl as an electrophile.
Yield=64%; ratio (5b)A/(5b)B=1/2.1; de=36%
1
(5b): H-NMR (l (ppm), CDCl3): 5.44 (s, 1H (5b)A,
O–CH–O), 5.42 (s, 1H (5b)B, O–CH–O), 5.41 (s, 1H
(5b)B, O–CH–O), 5.37 (s, 1H (5b)A, O–CH–O), 4.56
(m, 1H (5b)A+(5b)B, Cp), 4.27 (m, 3H (5b)A+(5b)B),
4.20–3.78 (m, 9H (5b)A+(5b)B), 3.60–3.25 (m, 10H
(5b)A+(5b)B, 2H–OCH3 and 2H–O–CH), 1.72 (m, 2H
(5b)A+(5b)B, –CH–), 1.45 (m, 2H (5b)A+(5b)B, –
CH–), 0.25 (s, 9H (5b)A, SiMe3), 0.23 (s, 9H (5b)B,
SiMe3).
(9b): Same physical data than for the product de-
scribed in ref. [13] except [h]D= −2.04 (CHCl3, c=
0.93): ee=25%; (R)-(7b) was the major enantiomer
([h]D= +8.17 (CHCl3, c=0.93) for (S)-(7b) [13]).
3.7. Hydrolysis of the bisacetal: synthesis of (7) and (8)
3.9. Synthesis of the i-oxa-trimethyleneferrocene (10b)
In a schlenk tube to a solution of bisacetal in
dichloromethane (ca. 0.15 mol l−1) was added 0.7 eq.
of paratoluenesulfonic acid monohydrate and ca. 130
eq. of deoxygenated water. The mixture was stirred 90
min at 60°C. After cooling, the solution was diluted
with ether, washed with water, dried on sodium sulfate
and evaporated to yield the dialdehyde which was
purified by flash chromatography on silicagel using a
cyclohexane/ether mixture.
Under argon, to a solution of diol (ca. 0.01 mol l−1
)
in 30 ml of freshly distilled benzene were added molec-
˚
ular sieves 4 A then 1 eq. of tosyl chloride. The mixture
was heated at 50°C. After 1.5 h the reaction was
completed and anhydrous sodium carbonate was added
after cooling. The solvent was evaporated and the crude
product was purified by flash chromatography on sil-
icagel to yield (10b) (68% yield).
1
(7a): Yield=83% H-NMR (l (ppm), CDCl3): 9.90
(s, 1H, CHO), 9.87 (s, 1H, CHO), 4.96 (dd, J=2.5 and
1.1 Hz, 1H, C5H3), 4.82 (m, 1H, C5H4), 4.79 (t, J=2.5
Hz, 1H, C5H3), 4.77 (m, 1H, C5H4), 4.59 (m, 1H,
C5H4), 4.52 (m, 1H, C5H4), 4.48 (m, 1H, C5H3), 1.50
(m, 6H, SnBu3), 1.29 (m, 6H, SnBu3), 1.05 (m, 6H,
SnBu3), 0.87 (t, J=7.1 Hz, 9H, SnBu3). 13C-NMR (l
(ppm), CDCl3): 194.50 (CHO), 192.88 (CHO), 84.89,
80.84, 79.60, 76.45, 76.36, 74.94, 74.02, 73.97, 70.61,
70.30, 29.11 (SnBu3), 27.30 (SnBu3), 13.65 (SnBu3),
10.81 (SnBu3).
(8b): Same physical data than for the product de-
scribed in ref. [13] except [h]D= −10.9 (CHCl3, c=
0.57) ee=ca. 30%; (R)-(8b) was the major enantiomer
([h]D= +36.2 (CHCl3, c=0.55) for (S)-(7b) [13]).
References
[1] (a) For a recent overview see: A. Togni, T. Hayashi, Ferrocenes,
VCH, Weinheim, 1995. (b) For a recent review on asymmetric
catalysis: S. Borman, Chem. Eng. News July 22 (1996) 38. (c)
6
(7b): Yield=92%. Same physical data than for the
product described in ref. [13].
For a recent review on non-linear optics: N. Long, Angew.
Chem. Int. Ed. Engl. 34 (1995) 21.
1
(8a): Yield=54%. H-NMR (l (ppm), CDCl3): 9.88
[2] (a) D. Marquarding, H. Klusaceck, G. Gokel, P. Hoffmann, I.
Ugi, J. Am. Chem. Soc., 92 (1970) 5389. (b) L.E. Batelle, R. Bau,
G. Gokel, I. Ugi, J. Am. Chem. Soc. 95 (1973) 482. (c) C.
Ganter, T. Wagner, Chem. Ber. 128 (1995) 1157.
(s, 2H, CHO), 4.86 (m, 2H, Cp), 4.70 (t, J=2.4 Hz,
2H, Cp), 4.55 (m, 2H, Cp), 1.75–0.75 (m, 54H, –
SnBu3).13C-NMR (l (ppm), CDCl3): 194.24 (CHO),

124
G. Iftime et al. / Journal of Organometallic Chemistry 565 (1998) 115–124
[9] G. Iftime, J.-C. Daran, E. Manoury, G.G.A. Balavoine, Angew.
[3] (a) O. Riant, O. Samuel, H.B. Kagan, J. Am. Chem. Soc. 115
Chem. Accepted.
(1993) 5835. (b) O. Riant, O. Samuel, T. Flessner, S. Taudien,
H.B. Kagan, J. Org. Chem. 62 (1997) 6733.
[4] (a) F. Rebie`re, O. Riant, L. Ricard, H. B. Kagan, Angew. Chem.
Int. Ed. Engl. 32 (1993), 568. (b) D.H. Hua, N.M. Lagneau, Y.
Chen, P.M. Robben, G. Clapham, P.D. Robinson, J. Org.
Chem. 61 (1996), 4508.
[5] (a) T. Sammakia, H.A. Latham, D.R. Schaad, J. Org. Chem. 60
(1995) 10. (b) Y. Nishibayashi, S. Uemura, Synlett (1995) 79. (c)
C.J. Richards, T. Damalidis, D.E. Hibbs, M.B. Hursthouse,
Synlett (1995) 74. (d) T. Sammakia, H.A. Latham, J. Org. Chem.
60 (1995) 6002. (e) T. Sammakia, H.A. Latham, J. Org. Chem.
61 (1996) 1629.
[6] (a) D. Price, N.S. Simpkins, Tetrahedron Lett. 36 (1995) 6135.
(b) M. Tsukazaki,, M. Tinkl, A. Roglans, B.J. Chapell, N.J.
Taylor, V. Snieckus, J. Am. Chem. Soc. 118 (1996) 685. (c) Y.
Nishibayashi, Y. Arikawa, K. Ohe, S. Uemura, J. Org. Chem. 61
(1996) 1172.
[7] (a) T. Hayashi, A. Yamamoto, M. Hojo, Y. Ito, J. Chem. Soc.
Chem. Commun. (1989) 495. (b) T. Hayashi, A. Yamamoto, M.
Hojo, K. Kishi, Y. Ito, E. Nishioka, H. Miura, K. Yanagi, J.
Organomet. Chem. 370 (1989) 129. (c) L. Schwink, P. Knochel,
Tetrahedron Lett. 37 (1996) 25.
[8] (a) J. Park, S. Lee, K.Y. Ahn, C.-W. Cho, Tetrahedron Lett. 36
(1995) 7263. (b) W. Zhang, Y. Adachi, T. Hirao, I. Ikeda,
Tetrahedron Asym. 7 (1996) 451. (c) J. Park, S. Lee, K.Y. Ahn,
C.-W. Cho, Tetrahedron Lett. 37 (1996) 6137.
[10] H. Jendralla, E. Paulus, Synlett (1997) 471.
[11] (a) W. Zhang, T. Hirao, I. Ikeda, Tetrahedron Lett. 37 (1996)
4545. (b) W. Zhang, T. Kida, Y. Nakatsuji, I. Ikeda, Tetrahe-
dron Lett. 37 (1996) 7995. (c) K.Y. Ahn, C.-W. Cho, J. Park, S.
Lee, Tetrahedron Asym. 8 (1997) 1179.
[12] H. Flack, Acta Crystallogr. A39 (1983) 876.
[13] G. Iftime, J.-C. Daran, E. Manoury, G.G.A. Balavoine,
Organometallics 15 (1996) 4808.
[14] Between two protons in a 1,2 position to each other, J=ca. 2.5
Hz; between two protons in a 1,3 position to each other, J=ca.
1.3 Hz from M.I. Levenberg, J.H. Richards, J. Am. Chem. Soc.
86 (1964) 2634.
[15] G.G.A. Balavoine, G. Doisneau, T. Fillebeen-Khan, J.
Organomet. Chem. 412 (1991) 381.
[16] A. Altomare, G. Cascarano, G. Giacovazzo, A. Guagliardi,
M.C. Burla, G. Polidori, M. Camalli, SIR92-a program for
automatic solution of crystal structures by direct methods, J.
Appl. Cryst. 27 (1994) 435.
[17] E. Prince, Mathematical Techniques in Crystallography,
Springer-Verlag, Berlin, 1982.
[18] D.J. Watkin, C.K. Prout, J.R. Carruthers, P.W. Betteridge,
CRYSTALS Issue 10, Chemical Crystallography Laboratory,
University of Oxford, Oxford, 1996.
[19] D.J. Watkin, C.K. Prout, L.J. Pearce, CAMERON Chemical
Crystallography Laboratory, University of Oxford, Oxford,
1996.
.
.